This shows you the differences between the selected revision and the current version of the page.
| control_algorithms 31/08/2010 10:41 | control_algorithms 23/10/2010 10:18 current | ||
|---|---|---|---|
| Line 12: | Line 12: | ||
| We decided to decompose the complex control problem into independent subproblems and solve them separately, if possible. This attempt resulted into the hierarchical control scheme, depicted in the first figure. Right at the start, we decided to completely separate the altitude control from the rest, essentially decomposing the 3D problem into separate 1D and 2D subproblems. Hence there is an independent //Altitude Control Layer// (ALCL), controlling the altitude of the vehicle, and a completely separated hierarchical structure, taking care of (2D) trajectory tracking (in terms of geographical position, i.e. latitude and longitude). Since both layers could be synchronized in time, this solution does not limit the ability of the system to track a complex 3D trajectory in any way. | We decided to decompose the complex control problem into independent subproblems and solve them separately, if possible. This attempt resulted into the hierarchical control scheme, depicted in the first figure. Right at the start, we decided to completely separate the altitude control from the rest, essentially decomposing the 3D problem into separate 1D and 2D subproblems. Hence there is an independent //Altitude Control Layer// (ALCL), controlling the altitude of the vehicle, and a completely separated hierarchical structure, taking care of (2D) trajectory tracking (in terms of geographical position, i.e. latitude and longitude). Since both layers could be synchronized in time, this solution does not limit the ability of the system to track a complex 3D trajectory in any way. | ||
| - | For a helicopter, this approach is very convenient, since the altitude might be controlled by a simple SISO PID controller driving the collective and throttle actuators (assuming that tilt, meaning roll and pitch, angles of the vehicle are reasonably limited and the vehicle is always in the upright position). Of course, this simplistic assumption excludes any kind of aerobatics, but it is sufficient for any "reasonable" flight envelope. There are interdependencies between the tilt angles, acceleration of the vehicle and collective settings, but can be safely neglected for reasonably limited tilt angles (by "reasonable" we mean 10°-15°). These neglected interactions can be viewed as an additional small disturbance, which can be safely handled by altitude and velocity controllers. They could also be compensated for by introducing some simple feed-forward branches between roll, pitch, velocity and altitude controllers, but it is disputable whether this solution would bring any measurable benefits. | + | For a helicopter, this approach is very convenient, since the altitude might be controlled by a simple SISO PID controller driving the collective and throttle actuators (assuming that tilt, meaning roll and pitch, angles of the vehicle are reasonably limited and the vehicle is always in the upright position). Of course, this simplistic assumption excludes any kind of aerobatics, but it is sufficient for any "reasonable" flight envelope. There are interdependencies between the tilt angles, acceleration of the vehicle and collective settings, but can be safely neglected for reasonably limited tilt angles (by "reasonable" we mean 20°-25°). These neglected interactions can be viewed as an additional small disturbance, which can be safely handled by altitude and velocity controllers. They could also be compensated for by introducing some simple feed-forward branches between roll, pitch, velocity and altitude controllers, but it is disputable whether this solution would bring any measurable benefits. |
| The "2D" control scheme is hierarchically structured, consisting of five separate layers. The first layer (//Angular Rate Control Layer//, ARCL) is responsible for the angular rates stabilization in the three vehicle axes. The second layer (//Attitude Control Layer//, ACL) serves for the attitude stabilization, while the third (//Velocity Control Layer//, VCL) controls the vertical and horizontal velocities of the vehicle. The fourth layer (//Position Control Layer//, PCL) is used to to stabilize the vehicle position in space. The uppermost layer (//Trajectory Tracking Layer//, TTL) is responsible for the vehicle guidance along a pre-set trajectory. | The "2D" control scheme is hierarchically structured, consisting of five separate layers. The first layer (//Angular Rate Control Layer//, ARCL) is responsible for the angular rates stabilization in the three vehicle axes. The second layer (//Attitude Control Layer//, ACL) serves for the attitude stabilization, while the third (//Velocity Control Layer//, VCL) controls the vertical and horizontal velocities of the vehicle. The fourth layer (//Position Control Layer//, PCL) is used to to stabilize the vehicle position in space. The uppermost layer (//Trajectory Tracking Layer//, TTL) is responsible for the vehicle guidance along a pre-set trajectory. | ||